Policy & Ethics

Why the Future of AI Music Might Hinge on an Aggregated Business Model for Licensing

Credit: Outlever

Key Points

  • The AI music industry faces a major roadblock in finding large-scale, ethically sourced training data.
  • Drew Silverstein, composer and President of AI Strategy at SourceAudio, argues that the solution is not better scraping technology, but a better business model built on partnership.
  • He champions an "ethical handshake" framework where rights holders have full control, opting into deals with complete transparency.
  • By aggregating catalogs from willing partners, this model creates a new marketplace and revenue stream for rights holders who previously couldn't compete at scale.
Drew Silverstein - President of AI Strategy | SourceAudio
When everyone chooses to shake hands freely and with full knowledge on something together, that's about as ethical as things can get.Drew Silverstein - President of AI Strategy | SourceAudio

Building a mutually beneficial system for AI models and music licensing has been a bumpy ride. A recent partnership between music licensing marketplace SourceAudio and voice AI leader ElevenLabs shows there is a promising way forward, but it won't be an easy road. The partners will need to navigate high-volume data needs, ethical issues around scraping the open web for music, and a general mistrust of AI use in creative outputs.

We spoke with Drew Silverstein, President of AI Strategy for SourceAudio, on his role in combating these challenges and what it means for rights holders. Having founded Amper Music, one of the first AI-driven music companies, combined with his experience as a composer, investor, and entrepreneur, Silverstein saw past the adversarial dynamics of AI and music to forge this new path, built on trust, mutual consent, and a framework that puts artists first.

  • The creator-first framework: SourceAudio's ecosystem is designed to give rights holders complete control, allowing them to choose which ElevenLabs deals are right for them. "Rights holders in our ecosystem are always opted out by default," Silverstein said. "We present them opportunities on a deal-by-deal basis with the option to opt in to that deal, supplying all the material terms they need to know to make an informed decision."
  • The ethical handshake: His philosophy is built on a simple but powerful vision of AI-powered music focused on informed consent. "While it appears simple on the surface, it's full of a million moving parts and is a really complicated process. But when everyone chooses to shake hands freely and with full knowledge on something together, that's about as ethical as things can get.."

For Silverstein, the roadblock to ethically training LLMs on music isn't a lack of willing partners, but a structural inability for any single rights holder to meet the monumental scale AI models require. The solution is not to find a bigger catalog, but to build a bigger table through aggregation.

  • The power of aggregation: To deliver on the volume of rights-cleared music needed to train AI models, SourceAudio aggregates, or packages, multiple creators' music together into one large offering. "If a licensee needs three million titles with specific criteria, and a rights holder's catalog has only ten thousand, they could never fill the order on their own. By aggregating folks together, we create opportunities that don't exist otherwise."

  • The new revenue stream: This method of aggregation for AI music is new, but a potentially lucrative avenue for rights holders. "As long as we can do it in a way that's consistent, using our internal guiding principles to do what's best for rights holders, then aggregation is a brand new revenue stream to explore," Silverstein said.

    Drew Silverstein - President of AI Strategy | SourceAudio
    By aggregating folks together, we create opportunities that don't exist otherwise. It's a brand new revenue stream to explore.Drew Silverstein - President of AI Strategy | SourceAudio

    Deep-seated skepticism of AI in creative industries poses yet another obstacle. But Silverstein explained his counterintuitive approach to open the minds of the naysayers. He argued the path to participation is not by force, but by radical transparency.

  • Education over mandates: For Silverstein, it's not about persuasion but about presenting the facts of how AI in the music industry can work. "Our job is not to convince them," he stated. "What I care about most is sharing the ground truth, the fundamental facts and realities, so we can all share that same understanding. From there, we'll have a conversation."

  • Opacity is the enemy: It's human nature to be afraid of what you don't understand. Silverstein and the SourceAudio/ElevenLabs partnership aims to make the AI music industry more transparent for creators. "If we don't understand something, our natural default is to avoid it or fear it," Silverstein explained. "But when they understand all the key elements, they want in."

SourceAudio's model of transparent, opt-in aggregation isn't just an ethical framework. It's creating a legitimate marketplace that stands in stark contrast to the data-scraping free-for-all or fair-use loopholes many rights holders fear. It proves that ethical practices and profitable opportunities are not mutually exclusive. Silverstein framed this moment not as an unprecedented crisis, but as the latest in a

the industry has to navigate.

  • Riding the wave: "Leaning into what's bleeding edge let's you catch the wave and benefit from the change," he argued. "Anyone is welcome to let a wave go. But the folks who lean in tend to thrive."

Ultimately when it comes to music licensing and AI, Silverstein has an easy litmus test to gut check if he's on the right path. "As a composer myself, I think about what would I feel proud of sharing with all of my friends and colleagues," he said. "If our business accomplishes that, then we are significantly contributing to the successful future of music."